_..-. ‘44..-1~‘..:- rm...‘ -...
`
`
`.1_‘5..»;f_-B‘EST:PQSSIBLE. copy .
`BEST POSSIBLE COPY
`
`l
`
`i
`
`r ’
`
`'
`
`i
`
`The potential carcinogenicity of modafinil was examined in 2 life-time bioassays one each in mice and
`rats. The drug was administered as an admixture in the diet for 78wks in the mouse and 104wks in the rat
`at 6, 30, and 60mg/kg/d doses for both studies. Basis for selection of these doses were13wk dietary tox
`studies. From both the rat and mouse studies. it is the opinon of the reviewer that the MTD was not
`reached. One of the reasons for this. is perhaps auto-induction of modafinil metabolism leading to below-
`adequate exposure levels. This enzyme induction was prominant with repeate dosing and has been
`observed in previous tox studies. The inadequate blood levels may have accounted for the minimal drug
`related effects in these 2 carcinogenicity studies. There was no drug related deaths (in mice and female
`rats). minimal clinical signs, no effect on food intake. hematology, clinical chemistry. or gross morphology.
`A sig decr was observed in mean wt of HDm rats (6.5% of cont) on wk80 and mean wt gain (47% of cont)
`wks 52-80. Mean wt gain was also reduced in MD&HDf rats (26% of cont in each gr) on wks 52-80.
`However, by the end of the study, mean wts and wt gains were comparable between all drug grs
`and the cont. There were no drug effects on B.wt or wt gain in mice. A sig and dose-dependent incr was
`observed in absol and rel wt of the liver in mice and histopath showed hypertrophy without necrosis. A
`slightly higher incidence of deaths noted in male rats due to increase in CPN. This is a common finding in
`aged rats usually males but the sponsor indicated that this finding was higher than the historical data and
`is therefore, drug related.
`it seems that many of the other lesions were secondary to CPN and not directly
`drug related.
`In both the rat and mouse studies, the type of neoplastic lesions were those common in
`aged animals (spontaneous) and the incidence was similar between cont and drug grs. Based on these
`results, it is the opinon of the reviewer that a conclusion can not be\on the potential
`tumorigenicity of modaf‘mil in rodents at this time.
`\‘mérxlé.
`
`Modafinil plasma levels were measured in these 2 carcinogenicity studies. Technical difficulties and small
`volume and other problems (see above) led to limited information. Mean plasma levels in mice measured
`only on wk4 were 102-402ng/ml in males and 23-205ng/ml in female mice and those in male rats 11-
`171nglml and in female rats 30-180ng/ml measured on wks 13. 52, 104. These values represent 0.002-
`0.04x in mice and 0.001-0.02x in rats, the maximum mean steady state cone of 11uglml reached in
`humans following 400mg dose.
`
`:
`
`‘
`
`~
`
`The sponsor recently (Apr 1996) submitted a 3mo tox and TK study in mice to further assess the MTD_
`selection for the carcinogenicity study. This study is reviewed and the reviewer concluded contrary to the
`sponsors conclusion. that an MTD was not reached. This is mainly based on lack of adequate exposure
`due to enzyme induction of modafinil metabolism with repeate administration.
`
`In the attached appendix is all the correspondance between the Division and the Sponsor regarding the
`adequacy of the carcinogenicity data, named in chronological order.
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`'
`
`I
`
`CC
`ioiv File/Orig mm—
` .
`[6. Fitzgerald/A AtrakchiIS. Hardeman
`
`10
`
`

`

`. . --.--.-. a--_.‘_a_.e_r__--_ r;-.
`
`‘éiés‘fPOS's‘IBLi: COPY.
`BEST POSSIBLE COPY
`
`' h
`
`Amendment# 08]
`IND#_
` .
`
`IND# _
` .
`Modafmil (Provigi1)/CEP1538
`Drug:
`Sponsor:
`Cephalon, Inc.,
`‘
`W.Chester, PA 19380
`Indication:
`Narcolepsy
`Category:
`Unknown mechanism, perhaps an «1 agonist
`Sub Date:
`Mar 31 97
`
`Apr 3rd 97
`Rec Date:
`Apr 24 97
`Rev Date:
`Aisar Atrakchi, Ph.D.
`Reviewer:
`Glenna Fitzgerald, Ph.D.
`Team Leader:
`Related IND/NDAs: N20-717
`
`The sponsor conducted an in vitro Morphological Transformation of BALE/3T3 mouse embryo cells to supplement
`the life-time mouse bioassay.
`
`This assay (GLP) was initiated in June 1996 and completed in Oct 96a—
` .
` .
`
`A preliminary cytotox assay was conducted at drug conc between 2-2000ug/m1 plus vehicle cont (DMSO) and
`incubated for 3d in -S9 and 4hr in +S9. Following exposure period, cells were washed and prepared; 10d after
`treatment, cytotox plates were fixed, stained with Giemsa, and scored for colony formation. The high conc for the
`main assay was selected to give about 20% survival; 3 other conc were also tested. Cytotox was assessed by cloning
`efficiency and rel CE. The main transformation assay followed standard procedures (Kakunaga, 1973; Heildelberger
`et al., 1983). The conc tested in the initial assay in -S9 were 150, 300, 600, and 1200ug/ml and those in +S9 were
`400, 800, 1200, and 1600ug/ml; a suspension (precipitate) was formed at 1200 and 1600ug/ml. In -S9 in the initial
`assay solubility hindered the use of conc >600ug/rnl and also high toxicity was seen, a repeate cytotox assay
`conducted concurrently with the main assay, tested the following conc: 38, 75, 150, and 300ug/ml. The positive com
`in -S9 was MNNG and dimethylnitrosamine in +S9. Transformed foci were spindle-shaped cells with dense,
`multilayered, basophilic in staining, random orientation, and invaded the monolayer. The results were expressed as
`dishes with transformed foci/total dishes, total transformed foci/total dishes, and transformation frequency (# of
`transformed foci per survivng cell). A positive response is the one where transformation frequency is sig incr over
`the negative com in a dose-response pattern (2 consecutive doses) including the HD. A sig incr at the HD only was
`considered equivocal. Any deviation from these qualifications will produce a neg response. A valid assay is that
`where the CE of the solvent cont is 225%, the no. of transformed foci will not exceed 3 foci/total (12-15) replicate
`dishes, and the positive cont should produce its expected incr in no. of foci.
`
`Results;
`The rel survival in the initial cytotox assay in -S9 ranged fi'om 0-94% and in +S9 from 6-106%. In the repeate assay,
`the rel survival in -S9 was 38-87% and in +S9 28-100% (38-300ug/ml cone). In the main assay, modafinil did not
`produce any transformed foci upto ISOungl in -S9 and only 1 foci out of 15 dishes (transformation frequency
`0.3 vs. <0.13 in cont) at 300ug/rnl. This was not statistically sig fi'om the solvent cont and within the historical data
`(provided in a table). In +S9, modafmil did induce any foci upto l600ug/ml. The positive controls produced the
`expected positive responses. It was concluded that modalinil in this in vitro assay and concentrations did not
`induce a statistically sig incr in the transformed cells in - or + $9.
`
`CC.
`[Div File/Orig IND-
` .
`IFitzgerald/S. Hardeman/A. Atrakchi
`
`
`
`

`

`f"
`
`BEST POSSIBLE COPY
`'BEST‘POS‘S’IBLE copv‘
`
`NDA #20-717
`
`Amendment
`
`Drug:
`Indication:
`Sponsor:
`
`Modafinil
`Narcolepsy
`Cephalon, Inc.,
`West Chester, PA 344-0065
`
`Corrsp Date:
`Rec Date:
`
`Jul 18 97
`Jul 18 97
`
`Rev Date:
`
`Jul 31 97
`
`Reviewer:
`
`Aisar H. Atrakchi, PhD
`
`erald, Ph.
`Team Leader: Glenna Fi
` .
`
`
`Related IND/NDA(s):
`
`.
`
`
`/S/
`
`
`Background:
`Apparently, this is an additional study in rats for the abuse potential of modafinil. Two studies,
`one in rat and another in Rhesus monkey, were submitted with the original IND (Jun 30, 1993).
`These 2 previous studies showed that modafinil was a re-inforcer for drug discrimination in
`monkeys (dose-dependent response at 0.03, 0.1, and 0.3mg/kg), and substituted for cocaine in
`drug discrimination tests in rats (21 50mg/kg).
`
`Submitted Study:
`Amphetamine-like Effects of Modafinil in Rats:
` .
`
` .
`
`This studyWW at the—and
`
` .
`
`in Jan 1996. This study is similar in design to the previously
`submitted to
`conducted rat study: rats trained to press lever for food reinforcement under a fixed ratio 32
`schedule during the daily 30min sessions. During training, d-amphetamine (lmg/kg) or saline
`were administered as controls to ensure that the rats were well trained in drug discrimination test.
`They were closed before the modafinil and amphetamine dose-response effects were done.
`‘ Modafinil was tested at 10, 30, 100, and 250mg/kg and amphetamine at 0.1, 0.3, 1.0, and
`
`3mg/kg.
`
`Modafinil at the highest dose of 250mg/kg, caused partial substitution in rats trained to
`discriminate d-arnphetamine from saline. However, the sponsor indicated that this dose was
`toxic since 1 rat died within 5hr of dosing and others exhibited low response rates at 224m
`postdose. At _>_100mg/kg no amphetamine-like effects were seen. This conclusion is similar to
`that mentioned and reviewed previously where modafinil substituted for cocaine at ZISOmg/kg
`in discrimination tests. Overall, it seems that modafinil re-inforces and substitute for drugs of
`abuse at doses 21 50mg/kg in rats & monkeys.
`
`cc.
`
`/Div File/Orig NDA# 20-717
`/G. Fitzgerald/A. Atrakchi/M. Malandrucco
`/S/
`
`“#3777
`
`
`
`

`

`“ "BEs'T 'P'o'S’SIBLE COPY '
`BEST POSSIBLE COPY
`
`Modafinil
`
`NDA# 20-717
`
`Cephalon Inc.,
`Sep 15 1997 Submission of additional data regarding the carcinogenicity potential of modafinil
`in mice.
`
`Rev Date: Sep 17 1997
`
`This submission contained:
`
`1.
`2.
`3.
`
`A protocol and results of dermal penetration study,
`Proposed protocol for a 26wk dermal study of modafmil in TG.AC transgenic mice and,
`Proposed protocol for 104wk mouse gavage car study in mice.
`
`Also mentioned by the sponsor, an ongoing oral dose-range finding study in FVB-N mice that
`has been extended to 13wks.
`
`.
`.
`. E [E _] I
`1
`.
`l
`E
`1.
`The FVP-N is the parent strain of TG.AC it is used here perhaps for economic reasons (less
`costly than the TG.AC altered variant) and other reasons not mentioned in the protocol. This was
`a pilot study therefore, non GLP.
`
`Methods;
`
`For the purpose of comparison, an oral dose was also studied. Modafinil was re ared in DMSO
` .
`at nominal cone of l8&108mg/ml for the dermal application (lOOul) and, in
`(oral
`suspending vehicle also used in the CD-l and OF-1 mice TK study previously submitted &
`reviewed), at nominal cone of 36mg/ml for the oral gavage (10ml/kg vol administered).
`
`There were 13933151552, for the dermal study, each mouse had lx2cm of their intrascapular skin
`region clipped free of hair 24hr pre-application. Doses for the dermal study were 60 &
`360mg/kg (note that 60mg/kg was the highest dose used in the 78wk mouse car study), and
`360mg/kg for the p.o. study. These doses were applied/administered once and blood collected
`after 1&4hr postdosing from 2-3/sex/dose. Modafim'l, modafinil acid (MA), and modafinil
`sulfone (MS), 2 major metabolites of modafinil, were measured using—
` .
`
`Emits;
`Modafinil plasma level following single dermal application incr with dose following 1&4hr
`measurements but the incr was non-linear (table below from the sponsor); same findings were
`obtained for the 2 main metabolites. Levels for modafinil and MA clearly declined after 4hr of
`application rel to the 1hr level, however, level of MS were generally higher for both sexes.
`Comparison of the same dose of 360mg/kg following dermal & oral dosing at 1hr postdose,
`showed that modafinil plasma levels afier dermal application were 21-26% lower than those afier
`p.o.; the same findings for the MA&MS. The mean plasma level of modafmil after single p.o.
`dose of 360mg/kg was lOSug/ml which is comparable to those reported eralier in the TK mouse
`
`1
`
`
`
`

`

`————i’
`BEST POSSIBLE COPY
`---‘ BEST"POSSIBLE'COPY '
`
`.
`
`z
`
`-
`
`-
`
`'
`
`(_
`
`Pilot Dermal study (Cont)
`
`study at 38ug/ml after lZOmg/kg (Cm analysis) in CD-l & OF-l mice. The same is true for the
`2 main metabolites, MA & MS.
`
`Table 1: Mean (Standard Deviation) plum communion: of modarmll and metabolites
`l andlhonnd‘tuaingkdemnl applicatiolud l honrpostonl administration
`
`
`
`'- mmammuuWMJWfimmmMmemm
`I‘D-Netball!“
`BQL-WW m¢<mm¢u
`
`A 4wk dermal range finding in FVB-N mice is planned to assess systemic exposure following
`repeate application and to select the proper doses for the 26wk dermal car study in TG.AC mice.
`The reviewer believes this is the same as the ongoing dose range finding that has been extended
`to 13wks.
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`WW5;
`
`Single dermal application of modafinil (in DMSO) at 60&360mg/kg to FVB-N mice-showed a
`reasonable systemic exposure to modafinil and its 2 main metabolites, the acid and the sulfone.
`A concurrent administration of a single gavage dose of 360mg/kg was done for comparison to
`plasma levels afier dermal application and to levels previously measured in a l3wk TK study in
`CD] & OF-l mice. Modafinil and its metabolite levels were comparable in these studies, but
`levels were lower after the dermal application.
`
`It can be concluded that single dermal application of modafinil provided reasonable plasma drug
`levels indicating that the drug readily penetrates the skin and reaches the systemic circulation.
`Systemic exposure need to be determined following repeate dermal application and the
`sponsor plans to do so in the near future or the study is ongoing.
`
`

`

`.-
`
`L - .
`
`42.44:.LMBEsT’ POSSLBLECORY: .
`BEST POSSIBLE COPY
`
`.
`
`
`
` .
`addresse all the standard parameters in a sub-chronic tox stud exce t for few items (see below).
`TG.AC mice will be purchased amide: 7-10wks old and will
` .
`be acclimated for about 2wks. There will be 20mice/sex/dose, using 3 doses (drug prepared in
`10% isopropyl alcohol), a control, and a positive cont ('1‘PA 2.5ug, 3X/wk, this cpd will be
`dissolved in acetone). The doses will be selected based on a 4wk dermal dose-range finding
`study in the parent strain DVB-N and the Agency’s concurrence will be requested.
`
`Cflmmsnfilfimmmndadnnmmm
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`It is recommended that plasma drug levels be measured at 2 time points during the study:
`perhaps at 2wks and end of study. To do this, it is recommended that additional
`mice/satellite (both sexes) be included for the TK analysis.
`
`Summary table for mean wt changes should be included.
`
`Under organ wts, what was the basis for selecting these 5 organs to weigh and not others
`or more or less?
`
`Under histopath, it is unclear whether all tissues will be examined; please clarify.
`
`it is unclear what will the negative cont be, is it the vehicle which in this case is 10%
`isopropyl alcohol or untreated mice.
`
`w
`
`v
`
`3.
` .
`will be the performing lab. CD-1 mice will be purchased from
` .
`at 6wks old (not more than 8wks), weighs lS-35g, and will be
`acclimated for about lwk. There will be 50mice/sex/dose, administered 3 doses and a cont gr.
`
`'
`
`'
`
`Want;
`
`1.
`
`It is recommended that plasma drug levels be measured at 2 wks, 12 & 24months of the
`study. To do this, it is recommended that additional mice/satellite (both sexes) be
`included for the TK analysis. The reason for the 2wk, is that modafinil causes
`autoinduction therefore, an early time point is advisable.
`
`2.
`
`Summary table for mean wt changes should be included.
`
`
`
`

`

`"'"BEST'POSSIBLE COPY ‘
`BEST POSSIBLE COPY
`
`Protocol for the 2yr car study (Cont)
`
`3.
`
`Organ wts should be measured.
`
`Histopath from cont and high dose only will be examined as stated in the protocol.
`Although, it is recommended that histopath done on all drug grs and the cont, however,
`‘ based on the low tox profile ofthis drug,
`it may be acceptable to microscopically
`examine only the cont and high dose grs and any mouse that died in moribund state, or
`due to unscheduled death.
`
`.
`
`4.
`
`cc
`
`/Div File/NDA# 20-717
`
`IG. Fitz erald/A. Atrakchi/S. Hardeman
`/S/
`/S/
`
`t S/ APPEARS THIS WAY 0N ORIGINAL
`
`APPEARS THIS WAY ON ORIGINAL
`
`
`
`

`

`'-
`
`.-_.
`
`BEST POSSIBLE COPY
`' «fl-4:4!» -:“‘~'BEST POSSIBLE COPY
`
`Tam-1BR AUG
`
`51997
`
`I
`
`5
`
`.
`
`.
`
`IE
`
`.
`
`IE 1
`
`.
`
`Review of Carcinogenicity Data
`
`AUG
`
`5 [997
`
`DATE:
`
`MI
`
`20-717
`
`( ._
`
`ABELICANI:
`
`I
`
`Cephalon, Inc.
`
`W:
`
`PROVIGIL (modafinil) Tablets
`
`W: Volumes 8-14 and 17-20 of 33 Containing Data, Results, and
`Study Reports of the Mouse and Rat Studies, and One Volume Dated 08/07/96 Containing the
`Data Layout and Diskettes for these Studies.
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`

`

`W'....r E-.:- .‘
`
`-‘~-
`
`BEST POSSIBLE COPY
`BESTPOSSIBLE con "
`
`LBaskgtound
`
`Dr. Aisar Atrakchi (HFD- 120) requested from the Division of Biometrics I a statistical review of
`the rat and mouse studies data as well as an evaluation of the sponsor’ s findings.
`
`11. 1111331513151!
`
`II.a. Design
`
`The drug was studied via dietary administrationin Sprague-Dawley rats for 104 weeks with 50
`animals per group. There were two control groups receiving the diet only and three treated
`.~ .. groups per sex receiving the drug at 6, 30, and 60 mg/kg/day. Diet and water were available ad
`lib.The liver, kidneys, lungs and gross lesions were examined for all animals of both sexes, the
`testes, aorta, stomach, and parathyroids for all male animals. For the remaining tissues full
`histopathological examination was performed only on all control and all high dose animals and
`on the animals of the low and mid doses which were found dead or killed in extremis.
`
`,
`( _.
`
`There were an additional 20 animals per group which were sacrificed at week 52 as part of a
`toxicity study. The findings of these animals are not part of the evaluation ofthe tumon'genic
`potential ofmodafinil.
`
`II.b.Spnnsgfis_Analxs_es_Qf_ths_EaLS_tudy
`
`S
`
`.1! l'
`
`The sponsor used Kaplan-Meier survival probability estimates and log-rank tests for group
`comparisons in survival. High dose males had a slightly higher overall mortality than controls or
`other treated males with a statistically significant postitive trend (p<.05). Among the females the
`mortality experience of the treated animals was comparable.
`
`W T
`
`he sponsor performed statistical analyses for selected lesions at the request of the pathologist.
`For fatal or non-fatal tumors Peto’s methods were employed. When all lesions occurred during
`terminal sacrifice the Cochran-Amitage dose-response test and Fisher-Irwin Exact test with
`Bonferroni adjustment for pairwise comparisons were used. The sponsor stated that there was no
`evidence of a treatment-related difi‘erence in the incidence, type or distribution of neoplastic
`lesions.
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`

`

`'f' BEs‘T ISGSS'IBLE COPY ‘
`BEST POSSIBLE COPY
`
`new
`
`5.1!].
`
`The survival experience of the male rats showed a statistically significant trend at p=.05 when
`using an approximate test for unadjusted positive trend. The probabilities associated with the
`adjusted Cox test or the generalized Kruskal—Wallis test were not significant Similarly, the
`unadjusted comparison between combined controls and the high dose was significant at p<.03
`but the corresponding Cox or Kruskal-Wallis tests did not reach statistical significance. The
`mortality experience of the female rats did not display a significant positive trend nor did any
`pairwise comparisons reach statistical significance (Figures 1 and 2, Table l).
`
`lummllataAnalxsis
`
`In the analysis of tumor data this reviewer uses the methods described in the paper of Peto et al.
`(Guidelines for simple sensitive significancetest for carcinogenicefi‘ectsin long-term animal
`
`experiments,
`_. .
`-
`-
`'
`International Agency for Research against Cancer Monographs, Annex to Supplement, WI-IO,
`Geneva, 31 1-426, 1980) and the method of the exact permutation trend test developed by the
`Division of Biometrics. The following criteria for the levels of significance ensure a false
`positive rate of about ten percent for the trend tests of the usual two-species two-sexes studies:
`Tumors with less than 1.00% occurrence in the control group are considered rare and a positive
`trend test is statistically significant when it reaches a p-value ofS .025 (one-sided). Higher tumor
`occurrences in the control group are considered common for these animals and a positive trend is
`statistically significant when its p-value is less than .005 (one-sided). An approximate
`permutation trend test is used when fatal and incidental tumors of the same kind are combined
`and have overlapping time intervals. All tests are survival adjusted and treatment groups are
`weighted by the actual dose levels. For tissues where not all dose groups were firlly necropsied
`only pairwise comparisons between the high and control groups were performed. The
`significance criteria for pairwise comparisons are 05.05 for rare tumors and (15.01 for common
`tumors.
`
`In this study all tumors were classified not as fatal and incidental but only as ‘undetermined’. It
`is therefore impossible to do the proper statistical analysis. This reviewer chose to analyze any
`possible increase in tumor incidence rates with the prevalence method, i.e. the tumors were
`treated as incidental.
`
`VWrth the exception ofliver, kidneys, lungs, and gross lesions in either sex, and the testes, aorta,
`stomach and parathyroids in the males, the tissues of the terminally sacrificed low and mid dose
`animals were not examined histopathologically. Therefore, for these tissues a trend test can be
`performed, for other tissues a pairwise comparison between high dose and combined controls
`will be performed. There were no trends or pairwise comparisons for either the male or female
`animals which passed the criteria for statistical significance.
`
`
`
`

`

`’ ""
`
`‘ ‘
`
`‘B‘E‘s'r' Poss'I'BL‘E cop'v‘ ‘
`BEST POSSIBLE COPY
`
`‘
`
`11d. Malidimnflthakatfimdx
`
`As there are no statistically significant tumor trends among either the male or the female rats, this
`reviewer evaluated the validity ofthe study. For this, two questions need to be answered
`(Haseman, Statistical Issues in the Design, Analysis and Interpretation ofAnimal
`Carcinogenicity Studies,WWW, Vol 58, pp 385-392, 1984):
`
`(i ) Were enough animals exposed for a suflicient length oftime to allow for late developing
`tumors?
`
`(ii) Were the dose levels high enough to pose a reasonable tumor challenge in the animals?
`
`The following are some rules ofthumb as suggested by experts in the field: Haseman (Issues in
`Carcinogenicity Testing: Dose Selection,WWW, Vol 5, pp 66-78,
`1985) had found that on the average, approximately 50 % ofthe animals in the high dose group
`survived the two-year study. In a personal communication with Dr. Karl Lin ofI-IFD-715, he
`suggested that 50 % survival of the usual 50 initial animals in the high dose group between
`weeks 80-90 would be considered as a sumcient number and adequate exposure. Chu, Cueto, and
`Ward (Factors in the Evaluation of200 National Cancer Institute Carcinogen Bioassays, ltmmal
`WWW Vol 8, pp 251-280, 1981) proposed that “To be
`considered adequate, an experiment that has not shown a chemical to be carcinogenic should
`have groups of animals with greater than 50 % survival at one year”. From these sources, it
`appears that the proportions of survival at weeks 52, 80-90, and at two years are of interest in
`determining the adequacy of exposure and number of animals at risk.
`
`In determining the adequacy ofthe chosen dose levels, it is generally accepted that the high dose
`should be close to the MTD. Chu, Cueto, and Ward (1981) suggest:
`
`“A dose is considered adequate ifthere is a detectable weight loss ofup to 10 % in a
`(i)
`dosed group relative to the controls.”
`
`“The administered dose is also considered an MTD if dosed animals exhibit clinical signs
`(ii)
`or severe histopathologic toxic efi‘ects attributed to the chemical.”
`
`“In addition, doses are considered adequate ifthe dosed animals show a slightly increased
`(iii)
`mortality compared to the controls.”
`
`In another paper, Bart, Chu, and Tarone (Statistical Issues in Interpretation ofChronic Bioassay
`Tests for Carcinogenicity,WWW: 62, 957-974, 1979), stated that
`the mean body weight curves over the entire study period should be taken into consideration with
`the survival curves, when adequacy of dose levels is to be examined. In particular, “Usually, the
`comparison should be limited to the early weeks of a study when no or little mortality has yet
`occurred in any ofthe groups. Here a depression ofthe mean weight in the treated groups is a
`
`
`
`

`

`' W'=‘BESTPOSSIBLE COPY” '
`
`BEST POSSIBLE COPY
`
`-5
`
`indication that the treatment has been tested on levels at or approaching the MTD.”
`
`Survival at terminal sacrifice ranged fiom 46 (High dose) - 66 percent for the male rats and from
`56-60 percent for the female rats, satisfying the requirement of a suflicient number of animals
`being exposed for a suflicient length of time to manifest late developing tumors. The body
`weight gains data showed that the high dose animals gained somewhat less than did their
`respective controls. The following tables give the mean weight gains, the mean body weights,
`and their difi‘erentials at weeks 52 and 104:
`
`(,
`
`-:
`( ,
`
`Mean Body Weight Gains
`
` FEMALES
`
`MALES
`Difi'erence
`
`Weeks 0- 52
`
`Controls
`270.0 g
`
`High
`256.7 g
`
`
`
`
`Weeks 0- 52*
`
`Controls
`493.2 g
`
`High
`474.7 g
`
`
`
`Difl‘erence
`-18.50 g
`(3.8 %)
`
`
`
`
`* The results for 0-52 and 0-104 Weeks are approximates as they are composites of weight gains for 0-13, 13—26,
`26-52, 52-80, and 80-104 weeks.
`
`-13.3 g
`(4.9 %)
`g
`
`
`
`Weeks 0- 104‘
`
`Controls
`460.2 g
`
`High
`412.9 g
`
`Difl‘erence
`—47.3 g
`( 10.3%)
`
`Weeks 0-
`104
`
`Controls
`371.8 g
`
`High
`334.2 g
`
`
`
`Difl‘erence
`~37.6 g
`(10.1 %)
`
`
`
`Mean Body Weights
`
`92.4
`
`% Controls Week
`
`'
`
`%Controls
`99.2
`96.9
`
`The survival experience of the high dose male rats was somewhat lower than that of the controls
`indicating with the other criteria that the high dose was close to the MTD for these animals. The
`survival experience of high dose female rats was similar across all groups and does not
`contribute to establishing the high dose as the MTD. However, as body weight gain was up to ten
`percent less than that of the controls, the MTD may have been reached for these animals too. In
`addition, the acceptable survival of all animals support the conclusion of a valid study. The
`pharmacologist may confirm these findings by evaluating the clinical signs and severe
`histopathological efi‘ects of this drug on these animals.
`
`APPEARS THIS WAY ON ORIGINAL
`
`APPEARS THIS WAY ON ORIGINAL
`
`

`

`‘B‘Esr POSéfBL'E 'CoPY‘
`BEST POSSIBLE COPY
`
`111W
`
`III.a. chign
`
`This study was conducted in ICO:0F1_ mice for 78 weeks. For each sex there were
` .
`50 animals per group. There were two control groups which received diet only and three treated
`groups which received the drug in the diet at 6, 30, and 60 mg/kg/day. The diet and water was
`available ad lib. There were an additonal 15 animals per control and treatment group which were it
`used for blood sampling and were not part of any survival or tumor analyses. Microscopic
`examination of tissues was performed only on all gross abnormalities, masses and tissues of all
`high dose and control (both groups) animals, on all gross abnormalities, masses and tissues of all
`animals that died or were sacrificed prematurely, and on. all gross abnormalities, liver and masses
`fi'om the intermediate and low dose groups. That is, with the exception of the liver, tissues of the
`terminally sacrified low and mid dose animals were not microscopically examined unless gross
`abnormalities or masses had been noted
`
`H.b.Sp_Qns.Qr3£_Anal¥ses_Qf_th§_MQus§_Smdy
`
`5.1!].
`
`The sponsor presented cummulative survival in graphical and tabular form. It was concludd that
`the ‘distribution of mortality for each sex was not superior in the treated groups when compared
`with the respective controls.’
`
`Iunmnllatammns
`
`For neoplastic findings the sponsor reported that the number of animals with neoplasms, the
`number of animals with more than one primary neoplasm, and the number of benign and
`malignant tumors were comparable in all groups. The statistical analysis of neoplastic lesions did
`not show any postitive trends.
`
`Elam
`
`5
`
`.1! l'
`
`This reviewer verified the sponsor’s conclusion that there was no statistically significant
`increasing trend in mortality with dose for either sex (Figures 3 and 4, Table 2).
`
`IuanDaILAnalysis
`
`The sponsor investigated possible trends only when there were at least three tumor occurrences
`in one of the groups. With the exact permutation trend test used by this reviewer this restriction
`was not necessary and any possible positive trend with dose was tested. A trend test was only
`
`
`
`

`

`————-—-—-——*r
`BESTPOSSIBLE COPY
`BEST POSSIBLE COPY
`
`—
`
`7
`
`appropriate for any fatal tumor and for incidental liver tumors. None ofthese reached statistical
`significance. Pairwise comparisons between high dose and combined controls also did not yield
`any statistically significant findings.
`
`md.WW3:
`
`Before concluding that the mouse study showed no tumorigenic efi‘ect ofmodafinil the validity
`ofthe study needs to be determed following the statistical criteria outline above for the rat study.
`
`At the end ofthe study more than 50 % (54% HD - 68% Controls) of the male mice had died.
`Numerically, however, there were 23 HD animals available to manifest any late developing
`tumors. The mortality among the female mice ranged fi'om 60 - 70 % at the end of the study with
`the ED animals having the poorest experience. The 15 animals of the HD surviving 1 V2 years
`may be insufiicient to demonstrate any late developing tumors, but may, on the other hand,
`indicate that the high dose was close to the MTD.
`‘
`
`The sponsor provided only individual and mean body weights but not mean body weight gains.
`The latter could not be computed by this reviewer because the body weight data were not
`provided on diskette. Using only the mean body weights (see table below),
`it is not clear
`whether the observed small difi‘erences constitute a drug efl‘ect. It rather appears that the high
`dose may not have been suficiently close to the MTD.
`
`Mean Body Weights
`
`
`
`'
`
`% Controls Week
`
`Controls
`31.5
`46.5
`47.0
`
`'
`
`Controls
`23.0
`36.0
`37.5
`
`%Controls
`100.0
`100.0
`96.0
`
`From the statistical point of view, this study had an insuficient number offemales exposed to a
`drug level which may not have reached the MTD calling the validity ofthis study into question.
`However, the numeric increase in mortality with dose would point to the high dose being close to
`the MTD, but the mean body weight data are inconclusive as there was no difierence between
`controls and High dose animals for the first year. For the male mice there was a sufiicient
`number of animals surviving for 1 ‘/z years but the high dose did not seem to be close to the,
`MTD. The evaluation of any toxic and severe histopathological findings are left to the expertise
`of the pharmacologist .
`
`-
`
`W I
`
`n this 104 week rat study all tumors were classified as ‘undeterrnined’ making a proper analysis
`impossible. In additon, only selected tissues were examined for all animals limiting possible
`
`C
`
`

`

`W"
`
`. - .-... BEST‘POSS'IBLE'COPY '
`BEST POSSIBLE COPY
`
`(
`
`s
`
`trend tests to these tissues. As no statistically significant trends or pairwise comparisons in tumor
`incidence rates were found the validity of the study had to be assessed. It was found that for
`either sex a sufficient number of animals lived long enough to manifest any late developing
`tumors. For the male rats the conclusion was that the high dose was close to the MTD and
`therefore that a valid study had been conducted. For the female rats the body weight gain data~
`supported this conclusion as well, however, the drug seemed not to have any efl‘ect on the
`mortality of these animals. From a statistical point ofview it is not clear whether the MTD was
`reached for the female rats.
`
`'
`
`This mouse study was only 78 weeks long. Only a limited number of tissues were examined for
`all animals making the standard tumor analyses unsuitable. The less powerful pairwise
`comparison tests for tissues where low and mid doses were not fully necropsied did not find any
`statistically significant results and neither did any of the trend tests. In the evaluation of the
`validity of the mouse study this reviewer observed that there were an insuficient number of
`female mice remaining at week 78 to manifest late developing tumors challenging the validity of
`the study. The increased mortality of the high dose animals supported the notion that the MTD
`was reached. However, the lack of difl‘erential in weight gain for the first fiill year casts doubt on
`this conclusion. The male arm retained a suficient number of animals at the end of the study but
`the high dose did not appear to have reached the MTD. Overall this reviewer concluded that the
`validity of the mouse study is very questionable.
`
`/S/
`
` Roswitha E. Kelly
`Mathematical Statistician
`
`/S/
`
`Concur:
`
`/S/
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`-- ,
`
`!___________—_—_
`
`

`

`———___.i7
`BEST POSSIBLE COPY
`‘BEST POSSIBLE cow ‘
`
`" ..-....
`
`‘
`
`cc:Archiva1 NDA 20-717, PROVIGIL (modafinil) Tablets, Cephalon, Inc.
`I-IFD-lZO/Division File
`
`HFD-lZO/Dr. Atrakchi
`
`HFD-lZO/Dr. Fitz